Learn more about principles of experiment flow by signing up for our master workshop on Testing Business Ideas.
Running experiments over time to reduce uncertainty in new business ideas is deceptively difficult to do. We find that teams try the tactics of experimentation without first understanding the principles. I’ve spent a couple decades in the lean and agile communities, so the problem is not unique to experimentation. Teams quickly adopt the practices of agile and lean because they seem easy to replicate.
Doing so however can be dangerous.
For example, auto manufacturers tried to copy the tactics of Toyota without ever understanding the principles behind them. One of the more notorious scenarios was the copying of the “Andon Cord”. The cord was a tactic, which started off as a cord (now in some cases a button) that workers could activate to stop the production line. They did so to warn management of a serious defect or issue that could become much more costly to fix down the line. This was especially true when the defect was then surrounded by other car components later in the production process. It meant hours of lost productivity and labor to remove the components to address the core defect.
Competing auto manufacturers heard stories of this process and decided to implement in their production lines.
It was a disaster.
The employees at other auto manufacturers would see a defect, pull the andon cord, and then be reprimanded by their managers for stopping the line. The factory had a quota of cars to build and stopping the line meant they’d miss their numbers. Afterwards, employees would never pull the cord again for fear of being fired.
You see when the cord was pulled at Toyota, the manager would come over to help diagnose the issue together with the worker. This was a very different experience because they created the practice from an underlying principle of jidoka. Jidoka translated into english roughly means “automation with a human touch” and is baked into the very DNA of the company.
Toyota workers knew at a principle level they should:
- Detect the abnormality
- Stop the line
- Resolve cause of the problem with management
- Incorporate improvements into standard workflow
We’ve seen similar scenarios with experimentation tactics, when teams copy what other companies do without understanding the why behind them. Teams have long running experiments that never finish or yield insights, they get frustrated and then build whatever they wanted to in the first place.
They could benefit from understanding the principles of experiment flow.
1. Visualize your experiments
It can be very difficult to manage your experiments without visualizing them. Just write them down and make them visual at first. If you are in person they can be on a wall or if you are distributed in the Strategyzer app. Then create a simple board to help manage them so you’ll be aware when things are getting blocked.
Then begin to add and rank the experiments in the backlog column. The experiments you’ll be running first are near the top.
2. Limit experiments in progress
Don’t try to run too many experiments all at once, especially if you have a small team. You’ll end up with a bunch of never finished experiments and a frustrated team. Set the limit on your board and prevent more experiments from being started until the other ones have progressed.
3. Continuous experimentation
Once you’ve helped experiments to flow with your team, turning the insights into action, then revisit the way you are working. Perhaps you’ll want to be more granular in your work breakdowns or change your work in progress limits. Don’t be afraid to change some of the practices once you understand the principles.
For example, your board might change to something like this:
Once you understand the principles of experiment flow, then you’ll be better able to get the most out of the practices. You’ll avoid long running, incomplete experiments with a check the box mentality. Remember the point is to help you reduce uncertainty and the flow of experiments should help you achieve that. Good luck!
Learn more about principles of experiment flow by signing up for our master workshop on Testing Business Ideas here.